Scholarlink

<p>Since 1961, the East Asian Meiyu has undergone two decadal shifts in the 1980s and 1990s. This study compares the influence of the Pacific, Indian, and Atlantic Oceans on Meiyu during two periods: 1952–1980 and 1990–2022. By employing a deep auto-encoder-based index (AE), we investigate the decadal variations of Meiyu under global warming, which demonstrates superior performance compared to single sea surface temperature (SST) indices. In the earlier period (1952–1980), the relationship between tropical SSTs and Meiyu was weak, with a dominant southward Pacific-Japan (PJ) pattern. This resulted in a northwest-southeast dipole moisture transport pattern, leading to poor predictability of Meiyu. In contrast, extratropical SSTs served as more reliable predictors during this time. However, in the later period (1990–2022), the relationship between tropical SSTs and Meiyu strengthened significantly, transitioning to the traditional PJ pattern. This shift led to a marked improvement in the predictability of Meiyu, as tropical SST anomalies exerted a stronger influence on atmospheric circulation patterns. These findings are further validated by Liang-Kleeman information flow (LIF) analysis, which highlights the enhanced causal relationships between tropical SSTs and Meiyu in the latter period. The study underscores the importance of considering multi-oceanic interactions and advanced statistical methods to better understand and predict decadal changes in Meiyu under the influence of global warming.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Neural network-based climate index analysis of decadal variations in meiyu: sea surface temperature signals before and after the decadal shifts

  • 期刊论文

  • 发表日期：

    2025年8月21日

查看全文
  • Peiyi Fan,
  • Bicheng Huang,
  • Jie Yang,
  • Ting Zheng,
  • Guolin Feng,
  • Zhihai Zheng

摘要

Since 1961, the East Asian Meiyu has undergone two decadal shifts in the 1980s and 1990s. This study compares the influence of the Pacific, Indian, and Atlantic Oceans on Meiyu during two periods: 1952–1980 and 1990–2022. By employing a deep auto-encoder-based index (AE), we investigate the decadal variations of Meiyu under global warming, which demonstrates superior performance compared to single sea surface temperature (SST) indices. In the earlier period (1952–1980), the relationship between tropical SSTs and Meiyu was weak, with a dominant southward Pacific-Japan (PJ) pattern. This resulted in a northwest-southeast dipole moisture transport pattern, leading to poor predictability of Meiyu. In contrast, extratropical SSTs served as more reliable predictors during this time. However, in the later period (1990–2022), the relationship between tropical SSTs and Meiyu strengthened significantly, transitioning to the traditional PJ pattern. This shift led to a marked improvement in the predictability of Meiyu, as tropical SST anomalies exerted a stronger influence on atmospheric circulation patterns. These findings are further validated by Liang-Kleeman information flow (LIF) analysis, which highlights the enhanced causal relationships between tropical SSTs and Meiyu in the latter period. The study underscores the importance of considering multi-oceanic interactions and advanced statistical methods to better understand and predict decadal changes in Meiyu under the influence of global warming.